Implementing a reset policy during a sequential variation test of content
View Patent ↗A variation testing system for performing variation testing of content is disclosed. The variation testing system allows an experimenter to continuously monitor the results of the variation test. In order to ensure that the experimenter is provided with valid results, the variation testing system may reset the variation test.
1. A computer-implemented method for conducting a variation experiment on a web page, the method comprising:
receiving indications that the web page is provided to a plurality of client devices, the web page associated with a first variation of the web page and a second variation of the web page that is distinct from the first variation of the web page, and each of the plurality of client devices receive either the first variation of the web page or the second variation of the web page;
calculating a confidence interval indicative of a range of differences in conversion rates of user interactions between the first variation of the web page and the second variation of the web page; and
responsive to receiving each of a plurality of the indications:
determining a difference in conversion rate of user interactions between the first variation of the web page and the second variation of the web page;
comparing the determined difference in conversion rate to the confidence interval;
responsive to the determined difference in conversion rate not being within the confidence interval, resetting the variation experiment; and
responsive to the determined difference in conversion rate being within the confidence interval, allowing the variation experiment to continue.
2. The computer-implemented method of claim 1 , wherein the second variation of the web page includes one or more modifications to elements of the first variation of the web page.
3. The computer-implemented method of claim 2 , wherein the difference in conversion rate of user interactions indicates performance of the first variation of the web page with respect to the second variation of the web page, and wherein the performance indicates increased interaction with the modified elements of the first variation of the web page that are included in the second variation of the web page.
4. The computer-implemented method of claim 3 , determining a magnitude of a difference in conversion rate of user interactions between the first variation of the web page and the second variation of the web page that is required to deem the variation test conclusive.
5. The computer-implemented method of claim 1 , wherein resetting the variation experiment comprises:
responsive to the web page being provided to a second plurality of client devices since the variation experiment was reset:
calculating an updated confidence interval based on user interactions with the first variation of the web page and the second variation of the web page that occurred after the variation experiment was reset; and
calculating an updated magnitude of a difference in conversion rate of user interactions between the first variation of the web page and the second variation of the web page that is required to deem the variation test conclusive after the reset.
6. The computer-implemented method of claim 5 , further comprising:
receiving indications that the web page is provided to the second plurality of client devices;
responsive to receiving each indication:
determining a difference in conversion rate of the user interactions between the first variation of the web page and the second variation of the web page that occurred after the variation test was reset; and
comparing the difference in conversion rate to the updated confidence interval; and
determining whether to reset the variation experiment based on the comparison.
7. A computer program product comprising a non-transitory computer-readable storage medium including executable code for conducting a variation experiment on a web page, the code when executed by a computer processor causes the computer processor to perform steps comprising:
receiving indications that the web page is provided to a plurality of client devices, the web page associated with a first variation of the web page and a second variation of the web page that is distinct from the first variation of the web page, and each of the plurality of client devices receive either the first variation of the web page or the second variation of the web page;
calculating a confidence interval indicative of a range of differences in conversion rates of user interactions between the first variation of the web page and the second variation of the web page; and
responsive to receiving each of a plurality of the indications:
determining a difference in conversion rate of user interactions between the first variation of the web page and the second variation of the web page;
comparing the determined difference in conversion rate to the confidence interval;
responsive to the determined difference in conversion rate not being within the confidence interval, resetting the variation experiment; and
responsive to the determined difference in conversion rate being within the confidence interval, allowing the variation experiment to continue.
8. The computer program product of claim 7 , wherein the second variation of the web page includes one or more modifications to elements of the first variation of the web page.
9. The computer program product of claim 8 , wherein the difference in conversion rate of user interactions indicates performance of the first variation of the web page with respect to the second variation of the web page, and wherein the performance indicates increased interaction with the modified elements of the first variation of the web page that are included in the second variation of the web page.
10. The computer program product of claim 9 , determining a magnitude of a difference in conversion rate of user interactions between the first variation of the web page and the second variation of the web page that is required to deem the variation test conclusive.
11. The computer program product of claim 7 , wherein resetting the variation experiment comprises:
responsive to the web page being provided to a second plurality of client devices since the variation experiment was reset:
calculating an updated confidence interval based on user interactions with the first variation of the web page and the second variation of the web page that occurred after the variation experiment was reset; and
calculating an updated magnitude of a difference in conversion rate of user interactions between the first variation of the web page and the second variation of the web page that is required to deem the variation test conclusive after the reset.
12. The computer program product of claim 11 , wherein the code when executed by the computer processor causes the computer processor to perform further steps comprising:
receiving indications that the web page is provided to the second plurality of client devices;
responsive to receiving each indication:
determining a difference in conversion rate of the user interactions between the first variation of the web page and the second variation of the web page that occurred after the variation test was reset; and
comparing the difference in conversion rate to the updated confidence interval; and
determining whether to reset the variation experiment based on the comparison.
13. A computer system for conducting a variation experiment on a web page, the computer system comprising:
a computer processor; and
a non-transitory computer-readable storage medium including executable code, the code when executed by the computer processor causes the computer processor to perform steps comprising:
receiving indications that the web page is provided to a plurality of client devices, the web page associated with a first variation of the web page and a second variation of the web page that is distinct from the first variation of the web page, and each of the plurality of client devices receive either the first variation of the web page or the second variation of the web page;
calculating a confidence interval indicative of a range of differences in conversion rates of user interactions between the first variation of the web page and the second variation of the web page; and
responsive to receiving each of a plurality of the indications:
determining a difference in conversion rate of user interactions between the first variation of the web page and the second variation of the web page;
comparing the determined difference in conversion rate to the confidence interval;
responsive to the determined difference in conversion rate not being within the confidence interval, resetting the variation experiment; and
responsive to the determined difference in conversion rate being within the confidence interval, allowing the variation experiment to continue.
14. The computer system of claim 13 , wherein the second variation of the web page includes one or more modifications to elements of the first variation of the web page.
15. The computer system of claim 14 , wherein the difference in conversion rate of user interactions indicates performance of the first variation of the web page with respect to the second variation of the web page, and wherein the performance indicates increased interaction with the modified elements of the first variation of the web page that are included in the second variation of the web page.
16. The computer system of claim 15 , determining a magnitude of a difference in conversion rate of user interactions between the first variation of the web page and the second variation of the web page that is required to deem the variation test conclusive.